Prometheus 监控 Redis 集群的正确姿势

Prometheus 监控Redis的正确姿势(redis集群)

Prometheus 监控 Redis cluster，其实套路都是一样的，使用 exporter。

exporter 负责采集指标，通过 http 暴露给 Prometheus 拉取。granafa 则通过这些指标绘图展示数据。Prometheus 收集的数据还会根据你设置的告警规则判断是否要发送给 Alertmanager， Alertmanager 则要判断是否要发出告警。

Alertmanager 告警分为三个阶段

• Inactive 触发告警的规则会被发送到这来。
• Pending 你设置的等待时间，即规则里面的 for
• Firing 发送告警到邮件、钉钉之类的

扯远了，开始监控 Redis cluster

redis_exporter 监控 Redis cluster

监控什么应用，使用的相应的 exporter，可以在官网查到。EXPORTERS AND INTEGRATIONS

Redis 使用 redis_exporter ，链接：redis_exporter

支持 Redis 2.x - 5.x

安装及参数

下载地址

wget https://github.com/oliver006/redis_exporter/releases/download/v1.3.5/redis_exporter-v1.3.5.linux-amd64.tar.gz
tar zxvf redis_exporter-v1.3.5.linux-amd64.tar.gz
cd redis_exporter-v1.3.5.linux-amd64/
./redis_exporter <flags>

redis_exporter 支持的参数很多，对我们有用的就几个。

./redis_exporter --help
Usage of ./redis_exporter:
    -redis.addr string
    	Address of the Redis instance to scrape (default "redis://localhost:6379")
    -redis.password string
    	Password of the Redis instance to scrape
    -web.listen-address string
    	Address to listen on for web interface and telemetry. (default ":9121")

单实例 redis 监控

nohup ./redis_exporter -redis.addr 172.18.11.138:6379 -redis.password xxxxx &

Prometheus 添加单实例

  - job_name: redis_since
    static_configs:
    - targets: ['172.18.11.138:9121']

Redis 集群监控方案

这个挺费劲的，网上查了很多资料，大都是监控单实例的，就这个是集群的，偏偏他的集群是没密码的。

prometheus监控redis集群

我试过的方案：

以下两种都会提示认证失败

level=error msg="Redis INFO err: NOAUTH Authentication required."

方法一

nohup ./redis_exporter -redis.addr 172.18.11.139:7000 172.18.11.139:7001 172.18.11.140:7002 172.18.11.140:7003 172.18.11.141:7004 172.18.11.141:7005 -redis.password xxxxx &

方法二

nohup ./redis_exporter -redis.addr redis://h:Lcsmy.312==/@172.18.11.139:7000 redis://h:Lcsmy.312==/@172.18.11.139:7001 redis://h:Lcsmy.312==/@172.18.11.140:7002 redis://h:Lcsmy.312==/@172.18.11.140:7003 redis://h:Lcsmy.312==/@172.18.11.141:7004 redis://h:Lcsmy.312==/@172.18.11.141:7005 -redis.password xxxxx &

本来想采取最low 的方法，一个实例启一个 redis_exporter。这样子的话，集群那里很多语句都用不了，比如 cluster_slot_fail。放弃该方法

nohup ./redis_exporter -redis.addr 172.18.11.139:7000  -redis.password xxxxxx  -web.listen-address 172.18.11.139:9121 > /dev/null 2>&1 &
nohup ./redis_exporter -redis.addr 172.18.11.139:7001  -redis.password xxxxxx  -web.listen-address 172.18.11.139:9122 > /dev/null 2>&1 &
nohup ./redis_exporter -redis.addr 172.18.11.140:7002  -redis.password xxxxxx  -web.listen-address 172.18.11.139:9123 > /dev/null 2>&1 &
nohup ./redis_exporter -redis.addr 172.18.11.140:7003  -redis.password xxxxxx  -web.listen-address 172.18.11.139:9124 > /dev/null 2>&1 &
nohup ./redis_exporter -redis.addr 172.18.11.141:7004  -redis.password xxxxxx  -web.listen-address 172.18.11.139:9125 > /dev/null 2>&1 &
nohup ./redis_exporter -redis.addr 172.18.11.141:7005  -redis.password xxxxxx  -web.listen-address 172.18.11.139:9126 > /dev/null 2>&1 &

最后只好去 github 提 issue。用我的中国式英语和作者交流，终于明白了。。。其实官方文档已经写了。

scrape_configs:
  ## config for the multiple Redis targets that the exporter will scrape
  - job_name: 'redis_exporter_targets'
    static_configs:
      - targets:
        - redis://first-redis-host:6379
        - redis://second-redis-host:6379
        - redis://second-redis-host:6380
        - redis://second-redis-host:6381
    metrics_path: /scrape
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - source_labels: [__param_target]
        target_label: instance
      - target_label: __address__
        replacement: <<REDIS-EXPORTER-HOSTNAME>>:9121

  ## config for scraping the exporter itself
  - job_name: 'redis_exporter'
    static_configs:
      - targets:
        - <<REDIS-EXPORTER-HOSTNAME>>:9121

Redis 集群实际操作

启动 redis_exporter

nohup ./redis_exporter -redis.password xxxxx  &

重点

在 prometheus 里面如何配置：

  - job_name: 'redis_exporter_targets'
    static_configs:
      - targets:
        - redis://172.18.11.139:7000
        - redis://172.18.11.139:7001
        - redis://172.18.11.140:7002
        - redis://172.18.11.140:7003
        - redis://172.18.11.141:7004
        - redis://172.18.11.141:7005
    metrics_path: /scrape
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - source_labels: [__param_target]
        target_label: instance
      - target_label: __address__
        replacement: 172.18.11.139:9121
  ## config for scraping the exporter itself
  - job_name: 'redis_exporter'
    static_configs:
      - targets:
        - 172.18.11.139:9121

这样子就能采集到集群的数据了。但是日志里提示

time="2019-12-17T09:10:49+08:00" level=error msg="Couldn't connect to redis instance"

午休的时候突然想明白了，只要能连接到一个集群的一个节点，自然就能查询其他节点的指标了。于是启动命令改为：

nohup ./redis_exporter -redis.addr 172.18.11.141:7005  -redis.password xxxxx &

Prometheus 配置不变

送上几张图片：

告警规则

groups:
- name:  Redis
  rules:
    - alert: RedisDown
      expr: redis_up  == 0
      for: 5m
      labels:
        severity: error
      annotations:
        summary: "Redis down (instance {{ $labels.instance }})"
        description: "Redis 挂了啊，mmp\n  VALUE = {{ $value }}\n  LABELS: {{ $labels }}"
    - alert: MissingBackup
      expr: time() - redis_rdb_last_save_timestamp_seconds > 60 * 60 * 24
      for: 5m
      labels:
        severity: error
      annotations:
        summary: "Missing backup (instance {{ $labels.instance }})"
        description: "Redis has not been backuped for 24 hours\n  VALUE = {{ $value }}\n  LABELS: {{ $labels }}"
    - alert: OutOfMemory
      expr: redis_memory_used_bytes / redis_total_system_memory_bytes * 100 > 90
      for: 5m
      labels:
        severity: warning
      annotations:
        summary: "Out of memory (instance {{ $labels.instance }})"
        description: "Redis is running out of memory (> 90%)\n  VALUE = {{ $value }}\n  LABELS: {{ $labels }}"
    - alert: ReplicationBroken
      expr: delta(redis_connected_slaves[1m]) < 0
      for: 5m
      labels:
        severity: error
      annotations:
        summary: "Replication broken (instance {{ $labels.instance }})"
        description: "Redis instance lost a slave\n  VALUE = {{ $value }}\n  LABELS: {{ $labels }}"
    - alert: TooManyConnections
      expr: redis_connected_clients > 1000
      for: 5m
      labels:
        severity: warning
      annotations:
        summary: "Too many connections (instance {{ $labels.instance }})"
        description: "Redis instance has too many connections\n  VALUE = {{ $value }}\n  LABELS: {{ $labels }}"
    - alert: NotEnoughConnections
      expr: redis_connected_clients < 5
      for: 5m
      labels:
        severity: warning
      annotations:
        summary: "Not enough connections (instance {{ $labels.instance }})"
        description: "Redis instance should have more connections (> 5)\n  VALUE = {{ $value }}\n  LABELS: {{ $labels }}"
    - alert: RejectedConnections
      expr: increase(redis_rejected_connections_total[1m]) > 0
      for: 5m
      labels:
        severity: error
      annotations:
        summary: "Rejected connections (instance {{ $labels.instance }})"
        description: "Some connections to Redis has been rejected\n  VALUE = {{ $value }}\n  LABELS: {{ $labels }}"